What Ruby versions should we continue supporting?

[corincerami]

I figured I'd open an issue for this since it's probably the easiest way to track discussion on it. I've been trying to decide if I think it makes any sense for Resque to continue actively supporting some of the older EOL Rubies. I feel like 1.8.7 is so far past its prime that it's probably putting more of a burden on contributors and maintainers than any potential benefit of pushing upgrades to 1.8.7 users, so in my opinion, dropping 1.8.7 seems pretty reasonable.

1.9.3 is similarly EOL, but I'm not sure how large of a base there is of people still using it, and a similar point can be made about 2.0.x, which was EOL'd in February of this year.

So the main point of contention, I suppose, is do we keep updating Resque for EOL Rubies, or do we just allow those users to lock their version at 1.26.0 and move on to focusing on supported versions?

[rjhancock]

I'd suggest dropping support for any version of Ruby that is no longer being supported.

[steveklabnik]

Historically speaking, a lot of resque's value is in legacy. That doesn't mean that every Ruby till the dawn of time should be supported, but it'd be nice to try and figure out some way to figure out what Rubies resque users use.

I'd also recommend that moving to a 2.x is a good idea if 1.8 support is gonna be dropped.

[rjhancock]

Then I'd stop at 2.0. Granted I'm working on a project that still runs on 1.8.7, but it's slowly moving to 2.0 (then newer after that).

[corincerami]

Yeah, I'm certainly not trying to argue that Resque should only support latest and greatest. I agree Steve, that it would be great if we had some kind of stats on what Rubies Resque users are on.

[pboling]

I have a large resque installation on ruby-2.0.0, (hundreds of millions of jobs per month), and will likely be for another few months. I think stopping at 2.0.0 is reasonable (biased!).

[wjessop]

+1 for Ruby 2 minimum. Anything older is ancient.

edit and by minimum I mean, I'd be happy with a higher required version too. People on old Rubies can maintain their own fork just as people on ancient Rails have to.

[IAMALLWE]

🎹

…

On Tue, Apr 11, 2017 at 8:18 AM, Will Jessop ***@***.***> wrote: +1 for Ruby 2 minimum. Anything older is ancient. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1526 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AYfKQl62NoNh9-qpvJhULees1sHo-y9Kks5ru32hgaJpZM4K9HMI> .

[claui]

I’d love for Resque to continue supporting Ruby 2.0.0 and above.

A large number of macOS users is basically locked into Ruby 2.0.0 today as it’s baked into all recent macOS versions including the latest one, Sierra. Apple also made Ruby 2.0.0 undeletable via System Integrity Protection. The average end user hardly has a chance to install and manage a second Ruby version in parallel while keeping his or her sanity.

While nothing of the above is obviously Resque’s problem in any way, and I think it’s fair to argue that Apple is 100 % responsible for exposing their customers to an insecure legacy mess like this, it would still simply affect too many end users if Resque were to drop support for Ruby 2.0.0, despite the latter being hopelessly outdated.

[wjessop]

I'd be surprised if there were many people running system Ruby when rbenv is available and easy to install. From my Sierra machine:

$ ruby -v
ruby 2.3.1p112 (2016-04-26 revision 54768) [x86_64-darwin16]

The instructions are readily available out there and could just be linked to in the README in the same section as the minimum version of Ruby is discussed.